Identifying baseline characteristics of placebo responders versus nonresponders in randomized double-blind trials of refractory partial-onset seizures.

Abstract

In add-on studies of partial-onset seizures, the placebo response, defined as a 50% decrease from baseline in seizure frequency, ranges from 0-19%. Reasons for this significant difference between placebo groups in different trials are not given in the literature. This exploratory analysis was undertaken to compare the baseline characteristics of placebo responders and nonresponders, in an attempt to identify common features. The pooled statistical analysis was performed on the database for three pivotal studies of levetiracetam (n = 904). Using the 50% response definition, we found that 45.6% of placebo nonresponders were on one antiepileptic drug at baseline, compared with 69% of placebo responders. The difference in number of baseline antiepileptic drugs was almost statistically significant (p = 0.056). Placebo nonresponders also tended to have epilepsy for longer than responders. The mean age at onset of epilepsy was consistently different between placebo nonresponders and responders (15.2 versus 20.8 years, respectively; p = 0.019). These findings suggest that the placebo response is higher in patients with partial-onset seizures who are taking only one antiepileptic drug at baseline and have later onset and shorter duration of epilepsy than in patients on more than one antiepileptic drug at baseline with earlier onset and longer duration of epilepsy.